EP0618355B1 - Method and device for controlling of the operation of a combustion engine for an automotive vehicle - Google Patents

Description

The present invention relates to a method and a device for controlling the operation of an internal combustion engine of a motor vehicle, by controlling at least one parameter for controlling the operation of this engine.

These methods and these devices have been developed in the state of the art (GB-A-2 042 772) to ensure the elimination of longitudinal oscillations of a motor vehicle.

In fact, in the state of the art, processes and devices of this type are already known which allow corrections and adjustments to be made to the parameters controlling the operation of a motor, in order to eliminate the oscillations, by monitoring a parameter and by introducing into a control loop of this parameter, a constant, at instants determined as a function of the detection of the oscillations.

However, these methods and these devices have a certain number of drawbacks, in particular as regards the determination of the times of application of the correction, insofar as the latter does not occur permanently.

It is also known from French patent application No. 91 11 919 filed on September 27, 1991, and published on April 2, 1993 under No. 2,681,908, a method for correcting the control parameters of an internal combustion engine. which consists, starting from an estimating model of the engine torque and a model for each estimating speed ratio of the transmission, to develop a variable which is a linear combination of the first and second derivatives of the engine speed and represents the oscillations extracted of the engine speed, to apply a variable correction to this variable to determine the variation of the engine torque and from this variation, to determine the correction to be applied to the engine control parameter (s).

However, this method has a certain number of drawbacks, particularly in terms of the complexity of the modeling, the volume of real-time calculations necessary to obtain the correction and the fact that very little action is taken on the start of the oscillation. .

The object of the invention is therefore to solve these problems.

To this end, the subject of the invention is a method for controlling the operation of an internal combustion engine of a motor vehicle, by controlling at least one parameter for controlling the operation of this engine, characterized in that it consists in developing a filtered engine speed signal representative of the engine speed oscillations, and in controlling this filtered engine speed signal to 0, to dampen the engine speed oscillations, by calculating a correction to be made to the engine torque and determining of the corresponding correction to be applied to said engine operating control parameter, using a digital regulator whose parameters have been previously calculated from a motor torque estimator model into which the engine control parameters are introduced operation of the engine and a filtered transmission estimator model for each gearbox report.

According to another aspect, the invention also relates to a device for implementing the method as described above.

• la Fig.1 représente un schéma synoptique illustrant le fonctionnement d'un dispositif de contrôle selon l'invention; et
• les Fig.2 et 3 représentent des schémas de principe illustrant la régulation appliquée à un paramètre de contrôle du fonctionnement d'un moteur, dans un procédé selon l'invention.

The invention will be better understood with the aid of the description which follows, given solely by way of example and made with reference to the appended drawings, in which:

• Fig.1 shows a block diagram illustrating the operation of a control device according to the invention; and
• Figs 2 and 3 show block diagrams illustrating the regulation applied to a parameter for controlling the operation of an engine, in a method according to the invention.

In Fig.1, the engine of a motor vehicle is shown diagrammatically by the block designated by the general reference 1 and receives as input parameters for controlling its operation, designated for example by the references P1 and P2.

The output of this engine is illustrated by the speed N thereof.

It is understood that this regime N is a function of the various parameters P1 and P2 for controlling the operation of this engine.

As mentioned previously, the method and the device according to the invention are suitable for damping the longitudinal oscillations of the vehicle generated by the oscillations of the engine speed.

The method according to the invention consists in developing a filtered engine speed signal NF representative of the oscillations of this speed N and in controlling this filtered engine speed signal to 0, by calculating a correction to be made to the engine torque and determining the corresponding correction to be applied to an engine operation control parameter, using a digital regulator.

The filtered speed NF representative of the oscillations of the engine speed N is obtained by passing this speed through a filter designated by reference 2 in this figure, while the digital regulator is designated by reference 3, the output correction thereof being designated by the reference CP and being applied for example to the parameter P1 for controlling the operation of the motor.

It goes without saying that the filter 2 can be integrated into the regulator 3.

It will also be noted that this control parameter is for example the ignition advance. However, the correction can also be applied to other parameters for controlling the operation of the engine as will be described in more detail below.

The parameters of the digital regulator have been calculated beforehand from a model estimating the engine torque into which are introduced the parameters for controlling the operation of the engine and a filtered transmission estimator model for each gearbox report. .

The identification of the engine torque estimator model does not present any particular difficulties and is carried out for example in accordance with the teachings of French patent application No. 91 11 919 mentioned previously.

The identification of the filtered transmission estimator model is carried out on the basis of the average engine torque over each constant sampling period of the regulator and of a filtered engine speed obtained from an engine speed interpolated linearly between values of speed measured at two determined angular references of the engine, framing the sampling instant.

This model can be obtained as follows.

The first phase of identification of this model is a phase of analysis of the functioning of the vehicle engine. This vehicle is equipped with an ignition and injection control computer which does not make any corrections to the parameters controlling its operation.

These parameters then come directly from a static mapping and we say that the system works in open loop.

It is then necessary to move to a stabilized engine speed, that is to say without too much residual oscillation and to accelerate so as to cause a significant oscillation of the vehicle.

At each determined angular reference of the engine, for example located before each top dead center thereof, the intake pressure or the air flow in the engine, the engine speed, the ignition advance and the time since the start of the measurement to date the results.

These various measurements make it possible to reconstitute in deferred time the torque delivered by the engine and the moment when it is delivered thanks to dating.

The calculation of the engine torque by a pressure, speed, advance triplet is described in the patent application mentioned above.

However, the working period, that is to say the sampling period, being constant, there is no synchronization between the determined angular references and the sampling instants.

However, to identify the model, it is necessary to know the engine torque and the filtered speed at each sampling instant.

• a) couple : le couple utilisé est le couple moyen calculé sur chaque période d'échantillonnage;
• b) régime : le régime non filtré à l'instant n est le résultat d'une interpolation linéaire entre les valeurs de régime mesurées aux deux références angulaires déterminées du moteur encadrant l'instant d'échantillonnage, ce qui est suffisant car le régime varie lentement d'une référence angulaire à l'autre.

These signals are reconstructed as follows:

• a) torque: the torque used is the average torque calculated over each sampling period;
• b) speed: the speed not filtered at time n is the result of a linear interpolation between the speed values measured at the two angular references determined by the engine framing the sampling instant, which is sufficient because the speed varies slowly from one angular reference to another.

The filtered regime is then obtained from this interpolated regime by passing through a high-pass filter of order at least equal to two to eliminate both a continuous component and a constant variation so as not to correct the system during constant acceleration without oscillation.

It can therefore be seen that, from this engine torque and the filtered speed, it is possible to identify the model that best traces the output signal for the same input.

Software can be used to identify the model from this information.

Additional information concerning this identification can be found in "IDENTIFICATION AND ORDER OF SYSTEMS" by IOAN DORE LANDAU, Treaties of New Technologies, automatic series, Hermès edition, 1988.

The sampling period is chosen to have a number of samples per period greater than four and to be greater than the short duration of an engine cycle.

The regulator can be constituted by a regulator with a structure known as RST of which one will find a detailed description in the document mentioned above.

This regulator is used to dampen the engine speed oscillations by slaving the filtered engine speed signal to 0, as illustrated in Fig.2.

In this figure, U ₀ represents the contribution of torque due to the driver of the vehicle, that is to say the disturbance.

The filtered transmission estimator model is represented by a discrete transfer function B / A where B and A are polynomials determined as indicated above.

The R and S parameters of the regulator are also polynomials.

We can then see that the closed loop transfer function of this regulator, linking the disturbance U _o to the filtered regime NF, is: $$\frac{\text{NF}}{{\text{U}}_{\text{0}}}\text{=}\frac{\text{BS}}{\text{AS + BR}}$$

The damping and the natural frequencies of the system are determined by the denominator AS + BR.

dans laquelle les inconnues sont les coefficients de R et de S, cette équation conduisant en fait à un système d'équations linéaires.We then determine a target or ideal polynomial called P and we solve the polynomial equation called BEZOUT: $$\text{AS + BR = P}$$

in which the unknowns are the coefficients of R and S, this equation actually leading to a system of linear equations.

The ideal polynomial P can be obtained by modifying the polynomial A insofar as it is not necessary to modify the oscillation frequencies or the damping of the high frequencies.

The polynomial A can then be decomposed in the form of first and second order polynomials so as to show frequencies and dampings and the damping corresponding to the frequency to be damped is then increased to obtain a modified polynomial which can be used as polynomial P.

Conventional methods for evaluating and improving the robustness of the regulator can then be applied to optimize its ability to maintain its performance when the characteristics of the system to which it is applied vary.

• de différentes marges :
  • * marge de gain
  • * marge de phase
  • * marge de retard
  • * marge de module
  
  Ces marges sont calculées à partir du diagramme de NYQUIST.
• de deux fonctions de sensibilité :
  • * sensibilité à une perturbation appliquée sur la sortie (NF) :
    Fonction de transfert : $${\text{S}}_{\text{YP}}\text{(q-1) =}\frac{\text{1}}{{\text{1+H}}_{\text{BO}}}\text{=}\frac{{\text{A(q}}^{\text{-1}}{\text{).S(q}}^{\text{-1}}\text{)}}{{\text{P(q}}^{\text{-1}}\text{)}}$$
    
    où H_BO est la fonction de transfert en boucle ouverte.
  • * sensibilité à une perturbation appliquée sur la commande (U_o) :
    Fonction de transfert : $${\text{S}}_{\text{UP}}\text{(q-1) =}\frac{{\text{A(q}}^{\text{-1}}{\text{).R(q}}^{\text{-1}}\text{)}}{{\text{P(q}}^{\text{-1}}\text{)}}$$
    

The robustness assessment is carried out conventionally through:

• of different margins:
  • * profit margin
  • * phase margin
  • * margin of delay
  • * module margin
  
  These margins are calculated from the NYQUIST diagram.
• two sensitivity functions:
  • * sensitivity to a disturbance applied to the output (NF):
    Transfer function : $${\text{S}}_{\text{YP}}\text{(q-1) =}\frac{\text{1}}{{\text{1 + H}}_{\text{BO}}}\text{=}\frac{{\text{A (q}}^{\text{-1}}{\text{) .S (q}}^{\text{-1}}\text{)}}{{\text{P (q}}^{\text{-1}}\text{)}}$$
    
    where H _BO is the open loop transfer function.
  • * sensitivity to a disturbance applied to the control (U _o ):
    Transfer function : $${\text{S}}_{\text{UP}}\text{(q-1) =}\frac{{\text{A (q}}^{\text{-1}}{\text{) .R (q}}^{\text{-1}}\text{)}}{{\text{P (q}}^{\text{-1}}\text{)}}$$
    

The sensitivity of each transfer function is determined by drawing the corresponding BODE diagram.

When determining the regulator, it is necessary to reach a compromise between all these criteria in order to respect them as well as possible. For this, there are several means of action.

Indeed, it is possible to modify and add roots to the polynomial P (to better specify it) without increasing the degrees of R and S to a certain point. The advantage of this method is to improve the robustness of the regulator without complicating it.

However, it is also possible to add a polynomial Hr or Hs.

$${\text{S = H}}_{\text{s}}\text{.S'}$$

These polynomials are defined as follows: $${\text{R = H}}_{\text{r}}\text{.R '}$$

$${\text{S = H}}_{\text{s}}\text{.}$$

If for example we introduce a polynomial H _r , the identity of BEZOUT becomes: $${\text{AS + BH}}_{\text{r}}\text{.R '= P}$$

où les inconnues sont S et R'By setting B '= BH _r , the equation is written: $$\text{AS + B'.R '= P}$$

where the unknowns are S and R '

It is the same for H _s .

The fact of adding this polynomial makes it possible to impose a root in R (or in S). However, for each root added, the order of R and S increases and thus the regulator becomes more complex.

It will be appreciated that the regulator shown in FIG. 2 makes it possible to quickly dampen an oscillation of the vehicle, but does not make it possible to completely control the acceleration and in particular the amplitude of the first oscillation.

This is reflected in the fact that this regulator only allows to choose the denominator of the closed loop transfer function, the numerator BS being entirely subjected.

To solve this problem, it is possible to introduce into this regulator, a direct action control signal developed from the engine torque U _o , to control the speed of the vehicle's acceleration.

This then results in the introduction of a polynomial T in the regulator as illustrated in Fig.3.

The closed loop transfer function of the regulator is then: $$\frac{\text{NF}}{{\text{U}}_{\text{o}}}\text{=}\frac{\text{B}\left(\text{S + T}\right)}{\text{AS + BR}}$$

The introduction of the polynomial T therefore makes it possible to partially control the numerator of the transfer function and requires the estimation of the torque U _o which would be delivered by the motor if no correction was made.

The global recursive formula for calculating the correction is then: $$\text{ΔU =}\frac{\text{T}}{\text{S}}{\text{U}}_{\text{o}}\text{-}\frac{\text{R}}{\text{S}}\text{NF}$$

The correction to be made to the engine torque at time n is therefore given by the relation: $$\begin{gathered} {\text{ΔU (n) = - (<figure-callout id="1" label="S" filenames="imgf0001.png" state="{{state}}">S</figure-callout>}}_{\text{1}}{\text{.ΔU (n-1) -<figure-callout id="2" label="S" filenames="imgf0001.png" state="{{state}}">S</figure-callout>}}_{\text{2}}{\text{.ΔU (n-2) -...) + t}}_{\text{o}}{\text{.U}}_{\text{0}}{\text{(n) + <figure-callout id="1" label="t" filenames="imgf0001.png" state="{{state}}">t</figure-callout>}}_{\text{1}}{\text{.U}}_{\text{0}}\text{(n-1) + ...} \\ {\text{- (r}}_{\text{0}}{\text{.NF (n) + <figure-callout id="1" label="r" filenames="imgf0001.png" state="{{state}}">r</figure-callout>}}_{\text{1}}\text{.NF (n-1) + ...)} \end{gathered}$$

The choice of polynomial T is based on several criteria.

Indeed, the correction made by this polynomial must be zero if the torque U _o is constant in order to avoid a permanent correction which would degrade the efficiency of the motor.

This condition imposes that T has for root at least the value 1, zero static gain.

In addition, the roots of S + T must be arranged correctly inside the unit circle.

One possible solution for calculating the polynomial T is to minimize a quadratic criterion grouping together the contradictory notions of amplitude of oscillations on the one hand and control energy on the other hand.

Another possible solution is manual calibration on the vehicle.

This regulator then makes it possible to determine a correction to be made to one of the parameters for controlling the operation of the engine in order to obtain a torque correction and therefore a damping of the oscillations of the vehicle and a control of the pace of the acceleration of the latter. this.

This correction is applied for example thanks to the ignition advance, but it goes without saying of course that an air actuator such as a valve or a butterfly valve can also be used.

In the case where the request for torque correction is applied by the ignition advance, as in the patent application mentioned above, the request for torque correction by the regulator, intervening all the sampling periods and the application of the feed, intervening all the determined angular references, are not synchronous.

It is then necessary to convert the torque corrections into advance corrections all the sampling periods, the calculation of the actual advance depending on the operating parameters of the engine at the time of ignition, being made all the angular references. determined.

In addition, it is also necessary to know the correction actually applied during the previous periods, which can differ from what was asked because of saturation.

For this, three solutions can be retained and it is then considered that the correction applied during a sampling period is either the correction applied to the last angular reference determined, before the next sampling period, or the average of the corrections applied to each angular reference determined during the sampling period, ie the average of the corrections applied to each angular reference determined during the sampling period, weighted by the time of application of each correction.

It can therefore be seen that the method and the device according to the invention uses a correction based on an identified model relating the engine torque to the filtered speed. and calculated at a fixed period, applied to specific angular references.

In addition, the regulator does not have a triggering threshold and it is possible to control the rate of acceleration of the vehicle and to determine a good compromise in vehicle comfort of use.

The robustness constraints are taken into account when calculating the regulator.

Finally, the real-time computational load of the regulator is low.

Claims (6)

1. A method for controlling the operation of an internal combustion engine of an automotive vehicle, by controlling at least one parameter (P1) for controlling the operation of this engine (1), characterised in that it consists in producing a filtered engine speed (NF) signal which is representative of the oscillations of the engine speed (N), and in locking in this filtered engine speed (NF) signal at 0, in order to deaden the oscillations of the engine speed, by calculating an adjustment to be made to the engine torque and by determining the corresponding adjustment (CP) to be applied to the said parameter for controlling the operation of the engine with the aid of a digital regulator (3) with its parameters calculated in advance using an estimator model of the engine torque into which the parameters for controlling the operation of the engine and for a filtered transmission estimator model are entered, for each gearbox ratio.
2. A method according to Claim 1, characterised in that the digital regulator (3) is a RST-structure regulator which receives a direct action control signal produced using the engine torque (Uo) to control the speed of acceleration of the vehicle.
3. A method according to either Claim 1 or 2, characterised in that the regulator has a period of constant sampling.
4. A method according to Claim 3, characterised in that the estimator model (B/A) for filtered transmission is identified using the average engine speed at each sampling period of the regulator and a filtered speed obtained using a linearly interpolated speed between speed values measured at two determined angular references of the engine, which angular references surround the sampling moment.
5. A method according to Claim 4, characterised in that the filtered speed is obtained by passing of the interpolated speed into at least a second order high-pass filter.
6. A device for controlling the operation of an internal combustion engine of an automotive vehicle for implementing the method according to any one of the preceding claims with means for controlling at least one parameter (P1) for controlling the operation of this engine (1), characterised in that it comprises means (2) for producing a filtered engine speed signal representative of engine speed (N) oscillations, and a regulator (3) for locking in this filtered engine signal at 0, wherein this regulator comprises means for calculating an adjustment to be made to the engine torque and for determining the corresponding adjustment (CP) to be applied to the said parameter for control of operation of the engine, wherein the parameters of the regulator are calculated in advance using the estimator model of the engine torque into which the parameters for controlling the operation of the engine and for a transmission filtered estimator model are entered, for each gearbox ratio.

Images